Global spending on AI data centers reached over 300 billion dollars in 2025. That is sixty percent more than the previous year. According to market research firm Intersect360 Research, the market will continue to grow, but the pace will slow later this decade as the first limitations of the current investment model become apparent.
The research firm bases its conclusions on surveys of members of its HPC-AI Leadership Organization (HALO) and a study of 458 companies with annual revenue of at least $10 million. According to Intersect360, this provides a broad picture of investments in AI infrastructure, ranging from data centers and GPUs to storage, networks, and software.
The majority of the investments come from hyperscalers, HPCwire reports. Companies such as Google, Meta, Alibaba, and ByteDance, together with specialized AI cloud providers, accounted for approximately $216 billion. A year earlier, that figure was around $120 billion.
Analyst Antonia Maar of Intersect360 notes that while growth is lower than the exceptional peak seen in 2024, it remains impressive given the size of the market.
AI Is Replacing Traditional HPC
Intersect360 will now combine traditional HPC and enterprise AI into a single category. According to CEO Addison Snell, this reflects real-world practice: virtually all organizations investing in high-performance computing now combine it with AI projects.
The study shows that 94 percent of the organizations surveyed are spending money on AI, HPC, or both. Pure HPC environments without AI have become rare.
Although investments continue to rise, Intersect360 expects growth to slow starting at the end of this decade. For the total data center market, the firm forecasts an average annual growth rate of 7 percent through 2030. AI-accelerated systems are expected to grow by 11 percent per year.
According to Maar, hyperscalers will eventually run up against practical limits in terms of capacity, energy, and efficiency. Snell also raises the question of how the massive AI investments will ultimately pay for themselves. He does not expect a market collapse, but rather a period in which expectations and business models will be adjusted.
Cloud Gains Ground
Within the enterprise segment, AI remains the primary growth driver. HPC is still growing, but is losing ground relatively speaking because AI investments are increasing much faster.
In addition, Intersect360 sees a further shift toward cloud environments. Organizations often still opt for hybrid models due to cost and data sovereignty concerns, but vendors are increasingly focusing their innovations on hyperscale and cloud platforms.
For enterprise AI systems, the firm expects an increase from approximately $22.5 billion in 2025 to $27.5 billion in 2026. A peak of over $40 billion is projected around 2029.
Energy as a Hindrance
In the longer term, Intersect360 cites energy supply as the biggest challenge. Geopolitical tensions, data sovereignty, and new technologies such as quantum computing could also influence the market.
Nevertheless, the research firm remains optimistic about the future. Intersect360 expects AI to follow a trajectory similar to that of the internet around the turn of the century: first a period of overheating, followed by a correction, and then sustained growth. By 2040, annual investments in AI data centers could therefore reach more than $1 trillion.